Health Catalyst, Inc. (HCAT) Earnings Call Transcript & Summary

In the important announcement category, the bar will be open for the duration of dinner. So everyone can help themselves. It looks like we are mostly at full capacity here, which is great. So let me be the first to welcome everyone to our 10th Annual Healthcare Analytics Summit. We are -- no, Salt Lake City is not the normal stomping ground for this, Chris. So we're very grateful for all of you to have made the trip out to see us. I'm just going to make a couple of introductory remarks to get us started before I turn things over to Dan. And just kind of keying off of this slide, it's actually great to see a lot of familiar faces back at HAS for their second, third or even fourth time, I know for some of you. So for those of you who have not been to has before, there is a lot of content, both this evening, for the agenda, we're very excited with the presentation for this evening for you all, but then also for the next couple of days. So kind of keying off of the slide up here, we would certainly encourage all of you to attend the keynote speaker and breakout sessions over the next couple of days as well as the analytics and AI showcases. Beyond that, there are, thankfully, a lot of our clients and prospective clients here this week. So we would encourage all of you to, certainly to interact with them on an ad hoc basis. We also have one opportunity that I'll touch on in a moment, where it will be a little more of a formal opportunity for you to interact with our clients. And then, for any of you who I haven't heard from, we still have a few slots available tomorrow with CFO and IR representation. So happy to find a time slot. You can just reach out to me. I'm not going to go through all the details on this slide. You all have a print out with this detail as well as the next couple of slides. But I did just want to touch on, in addition to us being excited to have a lot of our technology and professional services leadership team to present here this evening, we also have Matt Kolb, who is the Executive Vice President and Chief Operating Officer at Carle Health, one of Health Catalyst's largest clients. So thank you, Matt, for being here. I think it will be a great opportunity for you all to hear from -- to hear from Matt, both during his presentation and then during the Q&A panel. And then maybe I'll just -- I'll wrap it here. There's -- again, you all have a print out of this as well as a digital copy that I sent across. But this is kind of roughly a recommended agenda for your time at HAS. The one item I did want to call out here is, immediately following this evening's dinner, there is the analytics showcase as well as the AI showcase that are taking place on the first floor of the Grand America, so of this hotel. And to what I just alluded to, that will be a good opportunity. That will be a number of our clients who are presenting their -- some of their success stories that they've realized with Health Catalyst over the last several years. So that should be a good opportunity. We'll also try to -- for as many folks from the Q&A panel as possible, we will try to have people hang out here. So if you want to stick around after and grab anyone, we'll do our best with that. So with that, I'm going to pause. I'm going to turn things over to Dan here in a couple of minutes. But before Dan comes up, we would just recommend you all just to please get started on dinner, on your salads as well as take a few minutes just to introduce yourselves to others at your table. So I will stop there, and thank you again, everyone, for being here. We really appreciate it.

There we go. Okay. Hopefully, you had a couple of minutes there to get to know the folks at your table. And as Adam mentioned, we really appreciate each of you being here. One of the hallmarks of the Healthcare Analytics Summit is lots and lots of content. And so we figured today should be no exception. So we've got lots and lots of content to share with you. We are going to try to focus the content in areas that we hope are really meaningful to you. And we recognize that so many of you that have traveled to Salt Lake, get to hear from me all the time. So I'm going to be brief in my remarks. We brought several others that you're going to be able to hear from and perhaps the most important voice that you'll be able to hear from is Matt Kolb and we've allocated the most time to him, and he'll present and then he'll be part of the panel where we can have a Q&A session afterwards. But we've also brought a number of other leaders here, to this presentation, so that you can hear them present some interesting examples of what we're doing on the technology side, on the services side as well as ask some questions during the Q&A panel, and then several of them can linger longer afterwards as well. So let me start with a few -- I'll touch the tops of the waves of a few themes at an overarching company level, and I'll highlight some of the things that other leaders will go into more detail and delve a little bit deeper into. So this is an interesting time for Health Catalyst, in that, we're coming up on our 5-year anniversary from when we went public as a company. This next summer will be 5 years since our IPO, and it's a natural time and opportunity for us to reflect back on the last 5 years and think forward to the next 5 years. And so we, as a leadership team and as a Board, have engaged in a really meaningful strategy process this last cycle that's been deeper and more significant than a normal annual planning process. And reflecting back a few metrics that we've compiled here on this first slide. First, as it relates to where we were just before we went public, to where we found ourselves at the end of last year. One of the things we talked about at the IPO was, we were a high-growth company. We would grow at 20-plus percent. From 2018 through 2023, our compound annual growth rate was 21%. And we do take very seriously when we make commitments, to keep those commitments. Another thing we talked about from a profitability perspective, when we went public was, we were going to be on a meaningful positive trajectory towards profitability. We talked about a specific run rate time line that we also met over the course of the last 5 years, and we're excited to have just shared guidance last week that would place our profitability at the midpoint of the guidance, more than 125% higher this year than last year. And we're excited to continue that trajectory over the next 5 years. We've expanded our team member base to about double what it was just before we went public. And we've had the opportunity to dramatically expand the number of clients that we work with, more than quintupled since -- before we went public as a company. It's been quite a journey and quite a 5 years. But one of the things we've done as part of the stepping back process is update the Health Catalyst way. And many of you have been with us actually since before the IPO and will recall on the roadshow we talked about the Health Catalyst way. We've just gone through an updating process of that. And I'll touch on the tops of the waves of that, the Health Catalyst way, the mission, the strategy, the operating principles and cultural attributes continue to be central to who we are and what we do and how we build relationships with clients. So the mission has stayed the same for the 16 years of the company's existence, to be the catalyst for massive, measurable, data-informed healthcare improvement. And we've continued to use the flywheel, though we've refined it in this latest iteration, to represent our company's strategy with each client each year, where we begin at the top of the flywheel, when our clients take a leap of faith that the 3 components of our solution, the data and analytics platform, the applications that we've build on top of that platform and the expertise that we provide, when combined with our clients' best effort, will produce measurable improvement. That's the mission of the company. And when that measurable improvement occurs, we see the trust build between us and our clients. Our clients choose to deepen their relationship with us, and we work across our ecosystem to share those learnings to strengthen the ecosystem's ability to measurably improve. And that flywheel, that improvement flywheel spins faster and faster as the years go on. You'll hear in a minute from Matt Kolb. He's a great example. We were reflecting also, this is the 10th Annual Healthcare Analytics Summit. And actually, the first one was even further back. It was 12 years ago, and Matt was there. He was with another organization, he was with Allina Health at that time, and he was one of our great presenters. And since then, he has continued to grow in his career, in his achievements, in his seniority and moved to Carle Health. We're so honored that as is the case in many cases, we have this opportunity to continue those relationships. And that is centrally enabled by this focus on the mission and the focus on engagement, which is at the center of our flywheel that with our team members and with our clients, we focus deeply on making sure that, that engagement occurs at extraordinary levels. And that enables us to realize massive measurable improvement. We continue to emphasize operating principles and cultural attributes at Health Catalyst that we ask every team member to follow. We've slightly updated some of these, but we still focus on 4 each, improvement, accountability, respect and transparency. And from a cultural attributes perspective, we focus on continuous learning, deep long-term commitment to the mission, humility and striving for excellence in every job, every function and everything that we do. We need to realize that excellence if we're going to achieve our mission and be the ideal partner to our clients. We've also codified 1 new slide in the Health Catalyst way, which we refer to as Health Catalyst governing principle of living by the golden rule. What we mean by that is that we strive in every interaction with every stakeholder to treat others as we would wish to be treated with kindness, with humility and with respect. And with remembrance that in every interaction with another individual, what we know about what's going on in someone's life is a tiny proportion of what's actually going on. When we remember that, keep that perspective, it helps us to be a better partner, a more understanding partner and to live the golden rule. So that's the foundation of everything we do at Health Catalyst. And it was an important renewal, an opportunity for us to update that. We involved every team member at Health Catalyst in that process. We involved the Board. And we've just recently approved that, and we're publishing that now. Another really important aspect of our strategy work together as a leadership team and as a Board is our stakeholder strategy. And on this slide, you see a one-slide summary of the approach that we're striving to take that, frankly, every long-term successful and sustainable company needs to come up with and deliver differentiated value in a balanced way across their stakeholders. And for us, we have 3 key stakeholders, team members, clients and of course shareholders. We want you to know, we have a number of shareholders, and we have a number of equity analysts that represent shareholders here this evening. We're grateful to spend this time with you. We care about you as an important stakeholder, and we've tried to be very thoughtful and deliberate and analytical in understanding how we can drive differentiated value for you. But we believe it actually starts with the other 2 stakeholder groups. First, with team members, we have 3 pillars to our team members stakeholder strategy and the way that we provide team members with differentiated value. And the 3 pillars are engagement, career growth, and competitive compensation. We have, within our operating plan, specific elements within that plan that enable us deliver against each of those pillars and to measure our success. And I'll share some examples of that here in just a minute. For clients and our focus today at our user group conference has been very much on this client, stakeholder strategy and differentiation. We have simplified and focused on 5 key areas of differentiated value. There are 3 use case areas and 2 infrastructural areas. The 3 use case areas are clinical improvement, revenue and cost improvement and ambulatory operation, and the 2 infrastructural elements are measures in registries and data and analytics. And you'll hear more about that from some of our other presenters, each of those 5 areas has meaningful technology that we develop and also has important services and expertise that we provide. Now for shareholders, we understand at this point in our company's maturation and existence that for the most part, the most important metric that will drive shareholder value for us is profitability moving forward. We want you as shareholders and equity analysts to know we're focused on that as we think forward over the next 5 years that we understand that will be an important metric for success. We want to continue that trajectory that I showed you about the last 5 years, for the next 5 years and find ways to enable our shareholders to benefit from that increased profitability. You've seen that in the way that we think about our growth trajectory as a company to make sure that we are growing profitably and meaningfully. That includes also being mindful of our technology business and its importance with its margin profile and its profitability to drive shareholder value as a primary mechanism of driving that value. And we'll continue to emphasize the many reasons why our services, expertise, including tech-enabled managed services, really complement and strengthen our technology business and also provide really meaningful value to our clients and allows us to do work that's very mission consistent, which team members really appreciate as well. Okay. So I'm going to touch on just a couple of elements in each of those 3 stakeholder groups. So the first is team members. Many of you who have attended our Healthcare Analytics Summit in the past know that we use third parties to measure our success and our capability around areas like engagement, we use the Gallup organization to measure. And these -- this data shows the last 2 measurement periods. We've measured for the last decade, every year, our relative engagement across a benchmark of millions of respondents in Gallup's database. We have consistently averaged somewhere between the 94th and the 99th percentile in engagement. The most recent survey, during a difficult period in the company's existence and in the industry's existence, dipped down a little bit. We went down from 97th to 94th percentile. We're still really grateful to have that high an engagement level and we're very focused as a leadership team on how we strengthen that engagement level. It starts at the top of the organization, and we're grateful for incredible servant leaders, and we've showcased 2 recent promotions. We at Health Catalyst love to promote from within and both Jason's promotion and Dan's promotion were promotions from within, where these 2 individuals have been with the company for a very long time. Jason and Dan are here today. Jason, do you want to just raise your hand. You'll spend a lot of time with Jason moving forward. So we're so grateful for his capability set, the 11 years that he has spent at Health Catalyst and how well prepared he is. And then Dan, where is Dan? There's Dan. Dan has only been with us for 12.5 years. So he's operated in every functional area of the company and really well prepared for this role as our Chief Operating Officer. And many of these other individuals that you see on this slide and some others as well are here tonight to visit with you and share their perspective. I'm just going to touch the top of the wave, so we can get to Matt, which is why you're all really here. And we'll share this information on our website. So you have this detail, and you'll hear us repeatedly focusing in these areas. So with our second stakeholder group, we have simplified and focused. There's a reason why we're seeing R&D leverage, for example, as we simplify in these 5 areas of technology and services, value to clients, clinical improvement, revenue and cost improvement, ambulatory operations at the use case layer, and then measures in registries and data and analytics at the infrastructure layer. We believe we have defensible, differentiated value that we can provide, both from a tech perspective and a services perspective. And we maniacally measure the improvement that we're enabling with each of our clients in each of these 5 areas. And you'll see here several really meaningful examples that we've really been celebrating today with our clients, including many of our clients receiving flywheel awards for the amazing clinical, cost, financial, operational improvements that they have realized in partnership with Health Catalyst. Now one of the most important areas of investment that's we're excited to share a little bit more with you about and you'll hear from our CTO and a few other tremendous technology leaders at Health Catalyst is our next-generation data and analytics platform. We're excited about the strategic technology architecture that enables a real leap forward in terms of what we're able to offer from a value proposition perspective to our clients. We spent about half of our sessions today, in our user group, focused on delving a little bit more deeply into the next-generation data platform. And we wanted you to have the benefit of hearing from some of our key leaders in terms of what this new next-generation data and analytics platform enables and to answer questions from you about how our clients are responding. We're really excited and striving to manage the significant demand among our clients and their desire to migrate onto this next-generation data platform. Let me end by looking forward. Last week, we shared the result of really meaningful strategy work, really meaningful board work, leadership teamwork as we think forward to the next several years. We are excited about what we see. We're grateful to have strong and enduring client relationships, where they enable us to continue to grow in the near term, to continue to progress as a company in the near term while also seeing onto the horizon in 2025 and 2028, a growing and strengthening overall business that allows us to keep our commitments to our clients and also produce really meaningful shareholder value within our technology portfolio and within our services portfolio. We're excited to be right where we are. We look forward to the future. And there's maybe no greater reason why we feel optimism for the future than the experience that we have working with clients like Matt Kolb. So with that, let me turn it to Matt as our next presenter.

All right. Great. Well, thank you, Dan. I always ask him to manage down what I'm going to say a little bit better than he does. So we'll make the best of it. And as Dan said, my name is Matt Kolb, and I'm privileged to serve as the Chief Operating Officer at Carle Health. Carle, I describe a little bit more about who we are in a moment. My goal for tonight, and it's been -- it's really -- it's great to be back here. I actually have not been to HAS since 2013, which Tom told me tonight wasn't even officially HAS. It was just a user group meeting with like 12 people. And so it's great to see everything that's come since then. So my goal is to offer just a little perspective for this group about maybe what it's like to be a client with Health Catalyst and then share how our partnership has evolved over the past, really, 5 years. And so I think as Dan touched on, I want to start, though, with talking about Carle's mission. And I think this is important for a couple of reasons. One is because this mission statement really is the filter by which we make decisions about who should our partners be? Where should our capital be allocated? What are our priorities for the organization? It really is about how do we be a trusted partner to our communities in all healthcare decisions, not just clinical, but financial, community benefit, equity, all the rest. And I think Dan touched on mission alignment, which has really, I think, been a strength and an important part of our partnership with Health Catalyst is they think about healthcare in the same way that Carle Health does. And when that alignment is there, there's a lot of co-development that can happen, a lot of performance improvement and I think it has been a really important part of our success. And the other thing I think is important to note about this mission statement is that it's probably not unlike a lot of the clients that Health Catalyst has. Health Catalyst serves a large group of not-for-profit healthcare organizations that have an accountability to their community to deliver access to high-quality care. That's their first commitment. And with that comes, I think, a lot of purpose and a lot of alignment with the mission of health catalyst, but it also comes with community boards and a culture that has to process decisions and think through things, and that was certainly part of our transition as we've gone through an evolution with Health Catalyst. And so I think just keep that in mind that the client base is going through a lot right now in healthcare in general. And maybe as a side plug for all the smart, financially savvy, strategic minds in this room, if you're not involved in a local healthcare board, please consider that. Because it is a difficult time in healthcare, and we do need as many bright minds as possible serving in those roles. So if you're not, I'd encourage you to find a way to do that. I think it would be really very rewarding. So just to give you a snapshot of who we are. Some of you may be familiar, some of you not. Carle Health is located in Central Illinois, about 2 hours south of Chicago. We really cover the middle part of the state, from the Indiana border over to Iowa. Carle Health, as you can see there some of the statistics, owns a health plan. So Health Alliance medical plans with about 250,000 members. We're a partner with the University of Illinois to start The Carle Illinois College of Medicine. The marketing team likes to say it's the first engineering-based medical school though, I think there's like 5 universities that lay claim to that. So I always say one of the first engineering-based medical schools, but that's been a big part of our care model as well. But really, Carle is not so much these stats. It's the 17,000 team members, right, that show up to take care of patients and deliver 4,000 babies a year and 125,000 emergency departments and support $30 million worth of charity care in the community. So that's a little bit who we are, but it's really the people that make it special. So I want to talk now a little bit about our path with Health Catalyst, which really began probably almost 5 years ago, maybe 4 years ago. As we thought at Carle about, we have some challenges in front of us, and this was prepandemic. So this was 2019 and healthcare was certainly in a different place at that time. But looking to the future, we were thinking about how do we start to develop better data capabilities to drive performance improvement. How do we take all of the various data sources that we have across our organization. For us, that included a health plan that included EPIC and EMR vendor and put that together in a meaningful way that's actionable for our team members and that we can translate and turn into outcomes. And so I think the shared mission statement is important too. As I said, we have a lot of great mission alignment between Carle and Health Catalyst. And one of the first things that our teams did as we came together in our most recent partnership with the TEMS partnership was really think about where do our missions lineup and how does Health Catalyst support Carle that's been really meaningful for us, and I know is meaningful for other Catalyst's clients as well. But as I mentioned, we were sort of staring up at this mountain of what do we do? How do we get better? Some of you that have been around Health Catalyst for a while probably have seen this slide. And I think back in 2019, to be quite honest, Carle was probably around a level 1, maybe not even there. We had a lot of fragmented data sets, different points of care solutions, but it wasn't aggregated, and we weren't putting it together in a way to tell a story. So back in 2019, we actually went through a pretty extensive RFP process nationally, sent out a number of those, conducted several on-site interviews and ultimately selected Health Catalyst as our partner in the first quarter of 2020. And then March of 2020 came, and we validated immediately the value of that partnership because I do think we would not have been able to pull together some of the analytics that we use to support deployment of resources, community response to COVID if it weren't for that partnership. And so in some respects, it was sort of a partnership born out of a fire, but it really brought the teams together and was pretty meaningful. Obviously, since then, healthcare has continued to be under strain. You all know that. You are analysts and involved in the industry, and I don't need to tell you about what healthcare has experienced since the pandemic. But -- it's been pretty destabilizing for a lot of organizations, affordability challenges continue across the board. Access to care and equity remains a challenge nationally for every healthcare system, exacerbate that by the labor challenges and shortages that healthcare providers have been facing and trying to just provide care. And it's not an easy time for many providers and frontline care delivery organizations in healthcare. We haven't seen and maybe Health Catalyst clients are an exception to this, but nationally, we haven't seen some of the traction on measurable clinical improvement in the way that we would like to, from a population standpoint. I think there's great examples where we've had good success. And I do think Catalyst's clients would fit well above the mean because of the work that this team does, but as an industry, we certainly still have opportunity. And yet, I think there's also a lot of reason for optimism and a lot of reason for hope as an organization. And I think the -- what we've tried to focus on at Carle is really how do we -- what are the organizations that thrive? What will be the key to success in the next 5, 10 years as we sort of emerge from this pandemic period and think about how do we serve our community better? We know we need to redesign how we operate, and we need to do that at speed. And that is not an easy thing for any industry to do. That is a hard thing for healthcare providers to do, but absolutely an imperative. We have to accelerate efficiency in whole new ways and think differently about our care models and our business models. Think about partnerships like TEMS, which I'll talk about here in a moment. Those are things that are not intuitive to many healthcare providers. And traditionally, I think healthcare providers have had a history of build it yourself. And I think we'll have to think boldly about new ways of doing that. And then redesigning care. I was actually last night with -- at Carle, we have 4 cardiac surgeons to cover about 250 miles, 3 hospitals, 4.5. We have a cardiac surgeon in his 60s who's hanging in there. But we just count him as a half. And we were in a room and talking about this challenge. Recruitment is really difficult. How do we solve this in a way that will meet the needs of the community? How do we concentrate services where we need to, to meet that need? And it was very interesting because the physicians around the table said, "Well, let's really dive into the data. And let's really understand who does it best? How do we know? What are the costs? Does it make sense to do this at a regional hospital or should we concentrate services as hard as it is for the community and find a way to get patients accessed in a timely fashion". And from that, we have several follow-ups that will be data-driven in making a really hard decision. And that was an example for me last night, knowing I'm going to be here tonight of how grateful I am for this partnership, because that conversation would be much less rich if not for the work that this team -- our team at Carle and the Health Catalyst team have done together over the past few years. So I won't spend a lot of time on some of these dynamics. But I think as we took a step back in 2020 and thought about how do we work differently. We also approached Health Catalyst and really wanted to understand their model. And I was actually on the plane today, I read an article and it pointed out that the majority of performance improvement initiatives that organizations fail, something like 70% or more, but when the organization follows a playbook and a model that actually works, the success rate is over 80%. And so that was really what we found in our partnership with Health Catalyst as we emerged from COVID and started thinking about how do we transform care? How do we do these things? How do we measure best practice? How do we analyze that? How do we bring our clinical teams together, appeal to the scientists and the physicians that are a part of Carle Health and lead to outcomes improvement and measure that. And that's, I think, been a really successful part of our partnership. And from that really emerged the opportunity for tech-enabled managed service, which we entered into about a year ago at this time. And I think it's important though, as I share that process, that was not simply moving Carle Health's analytics function from Carle to Health Catalyst. That was certainly a part of it, but it was really a broader partnership that included access to Health Catalyst's full suite of technology and tools and offerings. We obviously wanted the transition for team members to be a smooth one, but it also supported Carle as we were looking to grow. We developed a different level of partnership where we have some shared incentives for actual clinical outcomes that are very meaningful for us. And I will tell you, too, that I think the Catalyst team sometimes celebrates Carle's wins more than the Carle team does, which is great. That's how -- that's sort of my benchmark for this partnership is working when the Catalyst team is as engaged and excited to see Carle Health succeed as the Carle team is. So that's been a really important part of the partnership. And then we've had some real intentional programs too about how do we equip our team members to learn basic analytics and capabilities as much as possible themselves? And how do we teach them to be more nimble at making data-driven decisions? As I mentioned earlier, all of the care transformation and change that we will have to go through ultimately needs to be driven by data in fact and outcomes and not anecdote. And so equipping our teams to think that way has been a really important part of that partnership as well, and I think valuable. And maybe since I'm with this group, I'll just mention -- from a financial standpoint, part of the savings opportunity here, as we assess this at Carle Health, we've really thought about. Okay, if we're going to invest more, both literally and figuratively in our partnership with Health Catalyst, how do we make that work for Carle Health? And the way we make that work is by selecting a few key partners and really, as you all know and maybe some of you have worked with healthcare organizations, there's often many different solutions that are a part of the organization that have been brought in by different leaders at different times. And so by taking a real concerted look across the organization and say, we're going to lose those because we know Catalyst has a tool and we're going to move our teams to that. There was really an opportunity for us to generate some efficiencies, some consistency and some financial value and savings. The other thing for me, as the Chief Operating Officer, is that by having a multiyear, multilayered partnership, it becomes a very predictable operating expense model for Carle Health. And at a time that the industry is up and down and fluctuating, that predictability is a very valuable thing for our organization too. And I'll just mention, I had a professor once who taught negotiations. And his mantra was always you have to find a way to expand the pie. Don't give people -- don't let people fight over a fixed slice, expand the pie and everyone can win. And I think that, that's come back to me at different times throughout this partnership, because I think that's what we've been able to do. We've been able to take sort of fixed pieces that exist within Health Catalyst, within Carle and truly offered something different for our team members through this TEMs partnership. So our team members, Central Illinois is not an easy place to recruit really bright statistical minds and analytic team members. We have great team members there. But recruitment and retention was a real challenge for us. And so part of the opportunity by partnering with Catalyst is this is Catalyst's core business. This is what Catalyst does. And so the pipeline to talent, the ability to develop people is something that Carle would not have been able to do on our own. And I think as our team members made this transition, they've appreciated that professional development as well. And I think at the 1-year mark, 97%, 98% of the Carle team members that were a part of that transition are still at Health Catalyst. And I believe, and Dan showed the engagement scores earlier, I'm told that they had some of the highest engagement scores. So for me, that's a huge win, because our Board, our community saw this transition. They were supportive of it. They saw the vision, but it was a change, and it was a little bit of, unorthodox isn't quite the right word, but it was not something that Carle had historically done. And so there have been a lot of eyes on, can we make this work? And so feedback like that is Exhibit A that, yes, in fact, we can, and it really benefits our team members. I touched on some of the benefits for us as well already just from a financial standpoint, but then the pace at which we can drive clinical change within our organization at a time that we really need to, has been a huge benefit for Carle. And then I think from a Catalyst standpoint, having a long-term mission-aligned committed partner like Carle, that's, I think, the portfolio of clients that any business would aspire to. And so I think this partnership has really been a beneficial one. I think, for everybody that's been involved. A little trepidation at the outset for our team, what is this, which is natural. And the Catalyst team did a great job walking them through that. And so it's been really great to see that. And I'll just -- I'll close here. I think you can probably tell, I've been very pleased with the partnership that we've developed over the past 4 years, been really pleased with and grateful for the support to our teams as we went through the transition in the last year as folks move to Health Catalyst from Carle. And then very grateful too that we have Health Catalyst as a partner. And I know TJ and Jason and Dave, I think, are going to come talk about some new development. But at a time, as I said, that the industry is facing some challenges, but there are a lot of really great organizations that want to be a part of the solution, but they need help. And some of the innovation that I know will come from Health Catalyst is incredibly valuable to Carle Health because we wouldn't be able to do that on our own. And through this partnership, I think we will be one of those organizations that thrives even as we move through some of these challenging times. So I will be happy to answer any questions during the Q&A about any aspect of Carle or TEMS or the selection processes or the evolution. But I'd just appreciate a few minutes to share how meaningful this has been for our organization.

All right. Thank you, everyone. I'm Dave Ross. I'm the Chief Technology Officer at Health Catalyst, joined up here by T.J. I want to get -- I think we both need new [indiscernible]. So I'm going to just go right to the next slide. But no, I'm excited to be here today and talk to you about really our underpinnings. I mean I'm going to reorient us to this place mat or what we call the cake side internally at Health Catalyst and just talk about our area of focus around data and analytics infrastructure is never going to change. And at the last HAS, I think I was only a few weeks into the role. And I think my first win that I celebrated internally was -- and I believe it was at this same Investor Relations dinner. Dan was coming up to talk about our investments in the data analytics platform. And the one key point I wanted him to make was that we'll always be improving this. There's not a set level of -- we got to a point now we stop. I think we'll always be improving what we do from a data and analytics infrastructure perspective. And I said that to Dan and I just started the role maybe 2 weeks, and he gets up here and he says, the word for word, what I just was speaking in his ear and I said, "something is working," like I picked the right role. So yes, I'm happy to just talk about why we're doing it? What we're doing? T.J. is going to talk about some of the why and the overall themes and we'll even get into the kind of tops of the wave slides that Dan covered in terms of what the actual underlying architecture is. And of course, we'll be available to answer questions, address concerns, et cetera. The time that we have after. So this is actually one of my favorite slides that we have at Health Catalyst. I couldn't be more committed personally to this. We've always supported kind of this teach you how to fish and our clients can do great things with the platform. So we've always supported this highly open, highly customizable environment. And that's not going away. And I think our DNA, we want to provide an environment that allows that kind of exploration, that kind of what will we learn next with our data. And of course, bringing in new data sources, combining them with things that we haven't seen in other clients. I mean that's our lifeblood. What we've also found is that, that creates an environment that's difficult to maintain at times. And there's a balance here. And so for me, on one side, you've got this fully customized start from scratch, build whatever you want within Health Catalyst. And of course, we've got expertise, knowledge, healthcare domain experience, et cetera, to help you do that. And on the other side, it's just a completely out-of-the-box experience where you install Health Catalyst and you get a few key dashboards, KPIs and metrics. And we've tried to thread this needle for quite some time. I really feel like we've now found this winning position, which is -- we've created a platform that allows you to do that custom work, but you're doing it on top of the standards that we've built over the past 12 years of being in the healthcare data and analytics space. So it's kind of like, the short story is, standards support custom better than custom supports custom. And we've had clients that have just started from scratch with custom. And that hasn't always led them to success. So there's the total cost of ownership, how long do they spend maintaining their platform? And how easy is it for us to keep them up to date? How is it for us to bring new technology to bear when they need it the most, and they often obviously, and you heard from Matt, there's been intense cost pressure in healthcare. And so you can imagine, it's tough to make an investment in maintenance activities when you're really just -- you're trying to figure out where you spend your critical dollars. So this kind of middle position of delivering standards to our clients and allowing them to build custom solutions on top of that is really where I feel that we are best and I'm excited to kind of bring that DNA into the platform that we're building. So T.J. is going to talk about the 5 aims. I don't know, do you want, well step over here...

All right. Dave is a better clicker than I am. All right. So as Dave mentioned, we've got the 4 major portions of our platform, really what we're doing with our solutions. What we're doing with CDP and our new modern data analytics platform is really trying to enable those 2 core foundational components to really accelerate and remove all the barriers to those higher-level use case, more use case driven kind of solutions, to accelerate and remove some of the barriers to value. So we set out really what we're seeing in the industry as a whole is, the healthcare has really gotten very good at bringing in lots of disparate data sources, collecting that data, creating a lake where there's all kinds of data that everybody could possibly get access to. The challenge is, what do you do with all of that data once you get it? And how is it that you take that data and turn it into meaningful insights that actually drive outcomes, impact of the ability to deliver care to the patients, how do you make sure that you can keep the doors open, and doing that in a financially sustainable way? And how do you improve your operations and allow your providers and those providing the care to really have the most optimal performance that they can possibly have and feel engaged in what they're doing and doing the work they want to be doing as opposed to making sure they're checking the box to do reporting and things of that nature. So what we really kind of summarize, what it takes to bridge that gap is creating high-value data and high-value analytics. And we set out really with the platform for 5 major aims that we felt like we had to accomplish in order to make that happen. We want to be integrated. So we've got to take all of those data sources that are coming in. We've got to be able to bring them together, and integrate them in a way that is meaningful to the downstream use cases. We want to be intelligent, we want to be modern and scalable, we want it to be extensible and we want to be accessible, and we'll dig into a little bit of what that means. But before we dig into exactly how we went about solving those problems, I wanted to help you guys just kind of frame what some of the problems are that we see and what we sought to solve with each of these. So integrated, we talked about -- they've got all these silos. We brought this data in, but it's sitting in silos. And you can't really bring all that data together to actually perform the downstream analytics that you need. It's siloed across disparate systems and you can't get it in and data mesh is a big thing in the industry. But right now, the reality in most health systems is that mesh doesn't exist to get those silos broken down. Intelligent. You don't have quality in the data. It's not geared towards the use cases that you're trying to provide. So you've got data, but it doesn't really fit for purpose and how do you get -- how do you enhance that data? How do you get it to the point where you can actually start to drive outcomes? Modern. This is just the most basics of it. It can't scale to the performance into the amount of data that we're trying to process and the different varieties of data that we're trying to process. Extensible. This gets a little bit to, as Dave highlighted, that middle ground. There's lots of point solutions out there in the market that will go, solve very specific purposes. And then there's highly open and customized, but how do you bring those point solutions together to answer the higher level kind of problems that we need to solve in health care? So how is it that the quality of care that I'm providing impacts the overall cost. And as I improve those things, how does it improve the financial position of the health system to be able to better position themselves in the market to maintain a financially viable operating model as they go forward? And then accessible. It's -- we could have the best data asset in the world and you could integrate. You could accomplish all of those other things. But if it's locked behind proprietary tools or SQL or you've got to have data analysts and I'm your data analyst who will be able to make that data actionable, you still fail. And so that is what we're seeing as, even if you've got that data, it's still sitting in a way that you've got now a limited set of resources that can go, grab that data and make it actionable. And Matt, I love what you were talking about how your data informed decisions -- what we're seeing today is that a lot of times data is used to reinforce the decision that's already been made rather than being able to actually use the data to an informed decision that you need to make. And by making the data more accessible, we want to allow that -- allow us to flip that to the point where you're using data to actually drive the decision-making process. So diving a little deeper into these. So integrated, how are we fixing the integration problems. So one, we've got end-to-end technology integrations. We've brought together a platform with the most modern technology that's available to integrate those data sources, bring them all together, create a comprehensive longitudinal record that you can build those analytics use cases off of that drives our platform and then leveraging the data centers that are out there. So we're built off of the standards in CMS and HL-7 are driving. We've got interoperability standards -- support for interoperability standards, support for clinical decision support standards and support for clinical quality measurement standards. And then we're also mentioned the interoperability standards. We want to make sure that if that data is in there that we have fire resources, support for CCDAs, every possible mechanism that health systems would need, all the different standards that are out there. So what is so great about them is there's always more than one. But we support a standard transformation to those to make that data more accessible as we curate it. Next part is intelligent. So this really goes into how we support and take the years of experience that we have in driving clinical outcomes and improvement? And how do we embed that in the data pipelines that we have? So we're using Healthcare.AI, that we're using the standard data models to go in there and actually drive insights into the data to highlight when something has changed, that's a variation that makes it no longer fit for use. So we're doing that and we're enforcing those and increasing our ability to get that curated data set by taking the years of experience we have in generating templates around this the data that we need to drive this use case from the source, here's how we go and do that. We're fundamentally trying to accelerate how we do that through the use of GenAI to start to generate all those insights. And then as we get that in there, we're leveraging Healthcare.AI to make sure that the data stays in control and fit for purpose, and that we pick up if there's a variation that occurs before the end user does. And that ultimately develops the trust that you need to actually use data to make decisions because if you don't have the trust in the data, even if you got it in front of you, you're probably going to go with your gut versus what the data is telling you need to do. Next one is modern. This is -- obviously, we have invested heavily with a number of partners to bring together industry leaders in the space of data and analytics management and then take those capabilities and apply those to healthcare. So Databricks, Snowflake, Azure Data Lake services. Those are the fundamental tenets of our new architecture. This gives us elastic compute. It eliminates those variables -- those cost variables that were going in, that would restrict access to the data before. It also enables modern data ops and scalability associated with that. So we've got CI/CD, continuous integration. We have standard Git for source control, things like that. So really, pushing to the very best practices of data management. And then we have standards that we apply to that to help them maximize their investment in those technology and the adoption of new technologies as we move forward. And then last -- sorry, not last, extensible, getting confused in my slides here. Extensible. This gets -- really gets and again is that custom experience that we need. We have models that drive the standards and allow for our clients to get measures, to get the data in, be able to measure their improvement, and see how they're doing on that. But then also customize those things to the things that are unique to those organizations. So it's one thing to say I'm measuring my diabetic A1C compliance rate, for example. But then I want to dig into the variables that are unique to my organization that might be driving my performance in diabetic A1Cs and that's where we have to have that open extensible experience to be able to allow them to go do that, allow them to look at the variables that are unique to that organization, and we do that in a number of ways. One is by having standard mechanisms to deliver the standard content, allow standard ability to extend that in a way that doesn't break the standards. And then also, when we talk about extensible, it's not just keeping it locked in those silos, but then making it available so that as you want to build an ecosystem with other vendors where you've got innovation centers within the health system that we've got prebuilt APIs for software developers and things to access that curated data that we've created for them. And then finally, extensibility kind of leads into the accessibility. But first and foremost, we talked about enabling decision-making. And that means taking it out of the hands of your SQL developers and your data analysts and putting it into the hands of the business analysts and the people that are actually out trying to make decisions. So one of the things you're going to see in the demo is we're going to demo Pop Analyzer, which is our self-service tool that makes that accessible. So we'll see that later. The other piece of it is enabling self-service data science in use cases. So you've got Databricks engaged there, and we're leveraging the full capabilities of the latest and greatest in data science technology. We're also embedding Healthcare.AI in all of our applications and in all of our self-service tools so that help drive the right decision and decrease the variability of decision-making based on the data. And then finally, it's just integration with notebooks, so meeting our end users in a modern environment that is readily available in the industry. So it's more -- it decreases some of those barriers, should option by using the latest and greatest tools in the industry. All right. So with that, I'm going to turn it back over to Dave.

T.J. asked me to present the 2 driest and busiest slides in the entire presentation tonight. The next one is even better than this. But when I think about this slide, the thing that really stands out to me is that second bullet of proportional scale and what that means to us. And I think back to one of the first slides that Dan brought up, which was showing our client base and showing the proportion of clients that are DOS installs today, which was roughly 100 and our total client base, which was north of 600 or just around that. And when I think about proportional scale, I think about scaling down to those smaller, more point modular opportunities and building a platform that can actually get small and cost efficient for those opportunities and also scale up to the largest, most complex health systems in the world. So to me, that's that bullet that stands out here is what we've built is something that can proportionately scale both up and down way better than what we had before. And we want to reach all of those clients. It's not just about the 100 DOS clients that we have today. We have those 400 other opportunities to build a more meaningful relationship. So in terms of the capabilities and the vision, I mean, T.J. touched on a lot of this, but we absolutely have our underlying infrastructure. So today, that's predominantly an Azure-based environment for our data platform, but we have other workloads running on AWS and GCP. Longer term, we absolutely feel like we're going to be in a modular state where we can mix and match according to our clients' needs. From a services perspective, we know we need to do all the things that T.J. talked about. So connecting the data sources, normalizing, harmonizing, transforming the data, getting it stored and prepared such that you can analyze it and then access it and make it accessible to both you and your constituents. So we're really talking about just the different components of that. Some of the newer things for us are real-time APIs and the applications that we plan to layer on top of the next-generation data platform. Because today, we've obviously added a lot of point solutions and a lot of other small, more modular technologies to our stack. And not all of those have been able to just easily plug in to the DOS environment that we've had in the past. So it's been a huge kind of opportunity for us. And again, it's back to that kind of like smaller scale, is can we take a smaller, more modular solution like an [indiscernible] et cetera, and get a data platform underneath it that brings data in, normalizes and curates it and brings it into the standards that we're talking about here without a huge client lift. I think that's a huge part of where we want to be in the future. And even this year, I mean, it's a huge part of what we're doing this year. It's not just the big large complex academic health system migration from DOS to the new platform. It's also the much smaller point solution built on top of our next-generation data platform that's really transparent to the end user. And from a cost perspective, it's much lower, it's basically pennies on what we were able to do in the past. So now for the big one and our favorite. When I tend to walk through this slide, and again, I'm trying to read the room here. I know that not all of us are technologists, and there's a lot of things on the slide that may not make sense. But I think if you just think top to bottom, you think core capabilities that we are getting from our industry partners. And so that's first years, Microsoft, Databricks, Snowflake, but then there's also other players involved, Okta, et cetera. As we move down the stack, we get into more of what Health Catalyst brings to the table, which is healthcare-specific domain expertise, models and all the experience doing that at scale. So I think as you move north to south, you're getting into what we do best and what our differentiator is. As you move left to the right, that's really the flow of data through our platform. So we're ingesting from sources. And again, that's often batch sources, reporting systems, et cetera, the back ends of transactional systems, but also messaging and streams. We're bringing it in, landing it in a raw form. And so one of the things that this slide doesn't show is that we've really got 3 tiers and they are medallion based. So you'll often hear us talk about the raw tier which is known as bronze, the kind of first curation, which is a standardized and normalized data set, and we talk about that as silver. And then we talk about a highly curated and highly valuable data set and even calculated metrics, really, really nitty-gritty things like, and I just talked to a client earlier today who -- the way in which they calculate discharge has to do with their interpretation of what midnight is and whether or not the patient got discharged there tomorrow. That's highly specific, it's highly specific to a client, and that's something we deal with as we take the data from that silver to bronze tier. As you go further down, so the bottom of this slide is really about how we integrate with other data platforms. And we know that as much as we would like to be, we're not going to be the only data platform that's been installed as in use at many of our large clients. And so getting into an environment where we can inter-operate, especially that silver layer, it's right in the middle allows us to supply other data consumers with data that's been curated first in the Health Catalyst data platform, and that's unlocked a lot of opportunities compared to what we could do in the past with the last generation of our technology. And so again, I mean, out the exhaust -- as T.J. calls it, out the exhaust pipe on the right side, and we have the analytics that we do with the data that we've curated. And of course, they're self-service, of course, they're supplying that data to other business users and absolutely, they're supplying that data to applications that use it to drive value. So that's my best effort to walk me through the marchitecture slide as we call it. And the last one we've got for you is just kind of an example of what I was describing earlier in terms of how we fit in a broader ecosystem and broader enterprise data cloud, knowing that, yes, we would love to be the only environment that a client has to do data analytics, but we also feel like we can be part of a broader ecosystem at a large client in terms of both receiving data from source systems and then contributing it back to make those source systems more powerful. And if they've stood up multiple data platforms and we've come across clients actually do that, we actually can contribute instead of saying, well, you only can have a siloed experience with Health Catalyst. And we think of ourselves as both can be the hub when needed, but can also be a spoke when that's appropriate for our clients. So that's it on the crazy slides. T.J. is going to be the Pop Analyzer demo. Thank you very much.

Sounds great. You have the trade-off for giving him those slides as I got to do the tech demo. So I was so grateful when it wasn't a live tech demo. So -- all right, so as I mentioned before, we're going to dig into Pop Analyzer, real quickly on what Pop Analyzer is, it's a solution that allows access, nontechnical users access to the data, allows them to easily create patient populations through a drag-and-drop interface really makes all the data that we've curated in the system, even the raw data available to our end users so that they can -- nontechnical users can go and explore in an environment in the language that they're comfortable with. So rather than thinking about in terms of value sets and discrete terminology codes and all of that, they can talk about it in terms of I've got a diabetic who needs an A1c and we allow that facilitate them to get those answers in that way. Once they generate those, they can explore the data, and we'll walk you through how we do that, but you can actually generate the population and then explore the attributes of that population to get new ideas and new insights and start to understand the things that are driving the success or failure within an outcome you're trying to achieve within a population. And then finally, as it relates to the downstream use cases, you can actually, as you create those, you can publish and explore those populations, make them available to your care management solutions, your patient engagement solutions, data science, you can actually define those populations and then make them accessible for additional downstream use cases. All right. So Pop Analyzer. As I mentioned, this is really a way to increase analyst effectiveness, reduce the development time, decrease the cycle time to develop new insights and then make those available to other downstream constituents to actually allow you to build on those capabilities rather than having to reinvent that themselves. One word I want to drill in here to with analysts is it's not data analyst. It's not from an IT perspective, but it's the business analyst as well it's those individuals that are actually out in the health system looking for data to inform their actions, and that is the primary audience that we want to gear this towards. So I'm going to walk you through a few of the personas and the key actions that they take, and then I'll walk you through the experience for this. So for the creator, the person defining the population, looking to gain the insights, it starts by they go and define a population. The next step that they do then is they explore, they export and they analyze that population. And I'll walk through what that experience looks like. And then finally, they can actually dive into an individual chart. And this is where you can get -- go from that high-level population of, let me see what's going on. And now I want to understand what's going under an individual chart, you can walk all the way down to the individual patients and gain those -- that level of insights as you walk through it. And then finally, as they've gotten to that level of expertise and they've curated and got that defined patient population, they can organize, share it, manage it, and it allows the others to stand on top of the shoulders of the data that they've created. And then, as I mentioned, they publish that population back to the data warehouse, and that makes it available not just within the data warehouse, but actually to the downstream applications and other use cases that they want to drive. So what are the benefits of enabling all of that? Well, first, it dramatically reduces your analytics development time, both through meeting the end users in the language and the types of interfaces that they need to be able to not get it out of SQL. It also allows them to leverage previously built populations so that they can actually stand on the shoulders of work that they've done, and they don't have to reinvent the wheel every time. Unleashes analytics use case and the use of data throughout the organization. So puts the hands in the data in the hands of the people that actually need it in order to take the action, speeds up analysis and accuracy. So as you define it, you're not just getting the data and saying, "Oh, that's just great. I've got this solution. I don't know where the data came from. I can actually drill down into it. And the ability to drill down into it, even get into individual patients. Ultimately, what that does is it encourages the trust in the data and the understanding of where the data is coming from. And that trust then breeds them to the end users to feel comfortable taking action on that data. And then finally, it reduced the cost. So you don't have to -- rather than having to hire an army of data analysts to go produce those insights for the business users, you're actually putting the data in the hands of the users and increasing accessibility will also eliminating the need to have really expensive data and analytics users. The second thing we're going to walk through here is the key actions of an administrator.

I'm going to give, I just got a heads up that we're a little behind schedule. So I'd love administrators, but let's get...

We'll keep going. Yes. I love it. All right. So where am I? I'm looking at 2 screens here. Okay. So as we walk through that, and I'll walk you through what the administrator tab was. So population tab. So this is where we start. We go through and we actually define the populations and we allow users to create a new population. So you see up on the top, you create new populations. You see in here, and you can actually use prebuilt populations to go start and create a new population and see either added existing ones or create a new population. As we walk through, then once you selected your population, you pop open and this is where you actually define the characteristics of that population. So -- which you'll see here is you'll see a diabetic A1C -- diabetics with a hemoglobin A1C greater than 9. So we've got a filter for patients aged 18 to 75. But they have a present diagnosis of diabetes, and that there's a lab result for A1C or the A1Cs, hemoglobin A1Cs that is less than 9. And then what you see over on the far right-hand side is the total number of patients that are in that particular cohort. And you see the filters down on the far right-hand side that you can drag in to fill or refine and filter that population. I mentioned, once we've defined this, you want to see who that 6,719 patients are. This is where we can log in and actually look at the DOS charts, so this shows you all of the patients in that population and some of the characteristics about them. So you can start to understand who are the patients that are in here, what are their attributes and characteristics. And I guess that breeds to the trust overall that you've got the right patients and the right populations as you move forward. And then once you've gotten that population correctly identified, this is the admin tab. So this is how you then go about the governance related to these so you can publish it and make that population then accessible either to downstream applications or other end users to go then build and add to those populations based off the use cases they might want to have. So if I'm a diabetic A1C and I want to know I'm a user who wants to understand the cost related to those or who my high cost. I can take my diabetic A1C populations. I can ask and show me those that have a cost of care greater than $70,000 or whatever in the loss. And I can add that additional filter on that without having to go and find that population. So with that, how was that Adam -- did I do -- all right. I will turn it over to Jason Jones.

Thank you much. Yes, I have to stand by the mic. I think I've had a chance once before to speak with several of you, and we'll cover a topic, I think, from a little bit of a different lens than what we've talked about before. So we are going to speak about Healthcare.AI. And I think you've -- many of you have heard about Healthcare.AI before. The way that I try to think of it as this, especially now working more on the services side of the organization is a lot of people talk about efficiency, efficient and wrong is not efficient. It's just wrong. And it doesn't matter how much you drive the cost down on the wrong answer, it's still wrong. Healthcare.AI is all about how do we help people get to the correct answer consistently and transparently, that's what this is about. And how can we do that at scale such that we can essentially put in the pocket of every substantive decision maker in a healthcare organization approximately a masters trained statistician or data scientist. That is what Healthcare.AI is about, and how can we do that in an appropriately affordable way. So we won't go through all of the levels right now, but we do spend from basic business intelligence integration. So if you're working in a Tableau or a Power BI or a Qlik, and that's what you know how to do, whether you're the BI developer or the administrator trying to make a decision that you will be guided by the best of what we have to offer in the field of data science while you're working on the tool that you already know and understand. Up through custom predictive modeling, including increasingly helping clients figure out how to optimize out-of-the-box predictive models from other places like their electronic health record vendors, it's a big problem, right? There's no other technology that we would just turn on in healthcare and say, yes, yes, it will work exactly the same, whether you're in Carle Health in Illinois or you are in New York Presbyterian and the chances of it working identically the same or where I studied L.A. County Hospital, not so likely different patients, different populations, different staff to work with. And then to those really, really big levers that leaders have to pull how well is our care management working? Where should I invest more? Where do I want to think about using a third-party vendor? How am I going to manage my drug and device costs while still being able to maintain or exceed great quality, but realizing that I've got to manage real cost pressures. Those are hard decisions. Those are multimillion dollar decisions. And if you're wrong, you can't be efficient. So the example that some of -- oh my gosh, that threw me too. I'm sorry. So I'm glad you'll be able to download this. The example, which some of you have walked through in the past is on the screen now. And we use this one because we usually have about 4 to 5 minutes with people to ask them to get those correct answers off of that top chart and then ask people to get the correct answers off of the bottom chart because a line chart is about as basic as it gets in data analysis in any type of industry, but especially in healthcare. And we routinely find every time we've done it, that we get a 10x improvement in people's ability to get the correct answers. I wish I could tell you that, that meant that they got the correct answer, a 1000%. It's not. Sadly, most people can get the correct answers from the top chart, about 4% of the time or I should say 4% of people can get the correct answers. Where are we? Where are we going? Give me the isolated extreme points of variation and help me disentangle those from systemic sustained change. We're just not very good at human beings, although we all want to believe that we are, we're just not that good. And so how do we help people get the correct consistent and transparent answer so that everyone leaves the table with the same understanding, it's the correct understanding, and we take the appropriate action. The most recent change that we have made has been to add on natural language generation to this, and this has helped us improve that ability for people to derive the correct answers. And it's interesting, we'll hear in a moment about some of the GenAI stuff. And one of the areas that I was focused on and reached out for us, how is this going if we have discrete quantitative data to work with? And the answer is not so well. So one of the areas that we're working on, especially going forward is how -- what do you do if you're in an organization that has 300 different KPIs that K kind of gets lost in the performance indicator. But we have lots of clients with 300 KPIs across 10, 100, 1,000 different ways of slicing and dicing the organization. How do you figure out where to focus. This is a scale problem. And as we transition over to the services side and in particular, the technology-enabled managed services, I want to be really careful to know how is it that we can augment what's there, which is by being able to build out as a technology organization, some of these tools that when I was at Kaiser Permanente, stewarding a $300 million to $600 million a year care delivery IT budget, we could not build this out because we were not a technology company. So how do we keep that part and from Matt keep the Jessica Summers who has been at Carle Health for 25 years. It was her first job out of high school, and then she went to college and came back. The Sharon Andersons, who have been there for 38 years, started as a surgery tech became a nurse and is still there leading big projects today. Those people are how we take these correct, consistent and transparent answers and turn them into actual improvement. Why? Because people trust the Jessica's and Sharon's of the world in a way that they will never trust a chart, and we have to find ways to bring them together, that's, I think, what we're enabling on the tech side with things like Healthcare.AI and now as we go to the services side with the Jessica and Sharon's of the world. So thank you much.

All right. I'm going to make up some time here. So I want to talk a little bit about -- we've already talked about these 5 focus areas -- strategic focus areas for us going forward. And I'll just call out that within each one of these, there's a meaningful tech-enabled managed services opportunity. So essentially, any time you're striving for an improved outcome, you need people, processes and tools to affect that improved outcome. And when we can own the technology and the processes and the tools, we can be a catalyst or an accelerant to those improved outcomes many times. And so hence, we have meaningful tech-enabled managed services opportunities available within each of these strategic focus areas of ours. I won't spend a lot of time here. These are the challenges that we've encountered over the last 18 months. We've gone around the country, Dan and Kevin and I and many others visiting with our customers. There are many challenges that they face, but we've just sort of codified a few of these as we think about what problem are we solving when we offer tech-enabled managed services. We want to reduce the financial pressure that they're facing. We want to help them recruit and retain talent. Matt talked about that. They see that outsourcing is a common go-to solution, but they're hesitant to go there because the options that they have aren't great and they can be culture killers. So they're very reticent to go down that path, they're begging us to be an alternative to some of the options that are on the table for them. Workforce reductions don't seem appealing that puts meaningful progress at risk and improved outcomes and other initiatives that they're working on, but they realize that they can't be jack of all trades. They need to partner with trusted partners who can specialize and bring capabilities to them. So how do we solve this? We have a tech-enabled managed services model where we contractually build in savings by month 10 of our relationships. We take their costs and reduce it by often up to 15%, sometimes 25%, built into the contract for them by month 10. So it's guaranteed savings for them. We care for their employees. Matt talked a lot about the engagement scores that we measure with those team members that come to us. We are maniacally focused on their experience. We want to make sure they're having a fantastic engagement experience. And oftentimes, they score among the highest of our Gallup engagement scores when we take those on an annual basis. We build in service level agreements so that we ensure that as we're getting more efficient, we're not compromising on the value that we're generating for the customer. They can rely on us to actually up-level the contributions of those team members as they come over as they receive training and technologies that can help them be more productive. And then we can specialize, we can scale investments in very targeted areas for them. We can purchase and acquire technologies that they would never imagine to acquire or purchase on their own for very specific use cases that we're serving. And we love this because it's a win for the client as we just illustrated, those are all wins for the client, but it's also a win for catalyst in many ways. For us, it's a deeper, longer-term relationship. Matt mentioned that not only was this a rebadging of employees, but it's also a deepening and the broadening of their engagement and interaction with our technologies. So our technology investments expand and become more sticky with our customers as well as just the broader partnership. We are truly intertwined. It's a marriage relationship, right? You don't enter into these lightly. And these become meaningful long-term opportunities for us to provide immediate savings to them, but give us an opportunity to earn profitability over time as we effectively adopt and deploy our technology and become more and more efficient over time, leveraging that technology, all while providing a fantastic work experience and engagement for those team members. So that's all I'm going to say it's better, it's faster, it's cheaper, and I want to share with you, so Allie Coronis leads our chart abstraction TEMS engagements in our team, it's probably about 200 team members now, I want to say they have come to us through rebadging. And so she's going to share a little bit about how this is better, faster, cheaper.

All right. So as Dan said, I work with our chart abstractors. And I don't know how many of you know what chart abstraction is but our team is generally comprised of clinically trained abstractors, usually nurses, that now are spending their time reviewing charts and pulling out data elements that are very specific to what the registries tell us it should be. And while that's really what they spend some of their time doing, we're really focused on helping our client partners to maximize the value of their registry participation. So we're looking to decrease the cost and the burden of collecting the data and increase the value of the information that comes out of the registry to improve care. And how we do that is we have a team of analytics engineers and they extract data from the EMR and different data sources so that the abstractors aren't spending their time, hunting and gathering for things that we can get accurately and efficiently. They can spend their time on data elements that require their critical thinking and their clinical knowledge, and they can shift their time more to looking at the outcomes and offering improvement opportunities to their clinical and client stakeholders. So on this screen are just a few examples of some of the areas where we've reached these efficiencies. We measure everything. And so one of the ways that we measure this is in reduced time per case. So looking at how long it takes to abstract the chart prior to extracting information how much -- how long it takes afterwards. And so the numbers on here are showing the savings in average minutes per case. But for each of these registries, there are hundreds of thousands of cases every year. So by applying this, our team spent saves. I'm trying to remember how many 50 FTE. So a lot less time on manual abstraction. So the tech enablement is a part. We do focus on abstraction workflow, but we're really focused on the tech enablement and Health Catalyst acquired ARMUS and ERS in the past couple of years. And I don't think anyone was more excited than our abstraction team members. You would think or I kind of thought it would be easy to get data into these registries and then to pull data out, and that is not the case in general. So both ARMUS and ERS, they have a lot of interfaces, both from hemodynamic systems, pathology systems, radiology. And so it really decreases the time that our abstractors are spending, just getting the right patient population or just pulling data that comes from discrete fields. They also both have great reporting and analytic capabilities. So our team members are spending less time exporting data sets, putting them in Excel, do a pivot table, put them in a PowerPoint. They have great canned reports, but then also ad hoc reporting and customizable reports. So I think the one bullet I'd maybe point to here is reducing data abstraction and costs and turnaround by 15% or more. That's 1 example, and we often see that be a lot higher. So moving on to data and analytics.

And sorry, I should have mentioned, Jason, like Allie leads our TEMS for data and analytics. So he's going to share similarly within that SKU of TEMS how we're better, faster, cheaper.

Thanks. And this time, I'll click first. So here we go. Okay. So the first thing is often to reground ourselves in a success story. So this was a wonderful one, and thank you to Leslie and team. She's actually the person who helps us figure out how to quantify these wins with our clients. And this was around care management and how it was that we could identify there are so many different places where you could focus with care management. There's a lot of overlap in care management, believe it or not. If any of you have experienced this yourself, you know it's the rare patient who has only diabetes, only heart failure, only high blood pressure, only need surgery, almost always as all of these and care management is often about how it is that you figure out who best can talk to the patient when and deliver what kind of support. And this starts to add up really quickly when you have duplicate people reaching out at sub opportune times to improve care. And ultimately, this opportunity resulted in being able to save approximately $30 million, which is meaningful to many of our clients. The other one I'll mention comes from another client in kind of the middle of the country. INTEGRIS Health. We have worked with them for quite a while, and there's a whole -- how that relationship has changed over the years because it's been a pretty massive shift in how we think about Health Catalyst role in data and analytics at INTEGRIS. There was a time when we were sort of the feeder into a shadow data and analytics system and now have gone all the way to actually being the core analytic competency within INTEGRIS Health. And if you don't know them, look them up, they're based out of Oklahoma. They have their Chief Medical Officer, will be here. She is amazing. And here, we've spanned all the way from how do we help them get a more cost-efficient and effective analytic compute environment through how it is that we deliver very, very targeted variable executive compensation performance measures for them, and everywhere in between. And specifically, what we focused on here is how it is that we could reduce the cost of their core analytics work. This is also an area where we focused on being able to spread expertise from 1 organization to another. At Carle Health, we were blessed going in with a very stable and successful analytic team. INTEGRIS brought us in because they had not been able to actually recruit the talent that they were seeking at all. And so we were moving people from other places around Health Catalyst to be able to serve that need, which they have been unable to recruit for a couple of years. And I will now, in the spirit of keeping us moving, go to ambulatory.

Thanks. Ambulatory is a newer SKU within our tech-enabled managed services offering. We have 2 clients now where we are managing the operations of clinics by employing the staff that the nonclinical staff that perform those tasks on a daily basis. This is an example. And a representation of exactly the work that we're doing at those clients today and a framework that we use when we think about opportunities to improve in those clinics, reducing care gaps when patients present and maybe it's for a well check and there are 5 other things that they need to have at that visit, making sure that those things are happening when they need to happen. Payer incentives. Oftentimes, there are payer incentives embedded in contracts, that if you perform these certain things or achieve these certain metrics, there are additional bonus payments and making sure that we're optimizing performance against those contracts that we know what they are, what measures were working against and providers are best equipped to perform well against those network leakage, where we have a capability in-house to provide certain services, how do we ensure that referrals are being captured within our network to provide those services where it makes sense and where it's convenient and accessible to the patient and then reducing denials in the revenue cycle space. These tend to be very large areas of opportunity where we can help our customers, and we manage those services for them. So what I want to do now is just first double down on the message that tech enablement is such a critical and important part of managing these services in a way that is better, faster and cheaper. And one area that we're very excited about that we're just starting to get into is AI enablement, generative AI and specifically in chart abstraction where we can have an assistant, a bot assistant essentially that says if I were abstracting this case, this is the answer that I would come up with based on the model that we've trained and then a live abstractor can respond, yes, that seems accurate or no, and help to train that model. And Dave is going to provide a quick overview of some of that work.

Actually, we've got a special guest. So we decided as a last-minute improvement to bring the man, Dr. Dan. This is Dan Samarov, PhD, who's the VP of Data Science Solutions. And what you're about to see, he's going to present real hard work that he's done to deliver AI within chart abstraction to deliver, of course, better margins in our TEMS business. I mean we're putting the T in TEMS and so we saved the best for last. So Dan.

All right. I was really hoping to introduce myself, as Dave Ross. But unfortunately, I can't. And I know we're a little tight on time, so I'm going to move through this a little bit quickly, and I'm happy to answer questions afterwards. As it come is no surprise, the chart abstraction is hard. I'm going to highlight a couple of reasons. First is that you need to have clinical expertise. Second is that the registries all have their own definition. So as an abstractor you have to be familiar with the specific registry. And third, you have to be familiar with the EMR, if you want to be efficient. So taking that into account, what we wanted to do was to see if we could build a copilot of sorts that could support our abstractors through abstraction process. What we didn't want to do was to create something that would require a net new workflow for them. We wanted to be embedded within the processes that they already had in place. And on top of that, we wanted to ensure that the way that we built it reflected how they consume information and what they would need to trust that system. So in developing our models and our tools, what we put into place was something that didn't just provide an answer but also provided supporting evidence from the clinical notes in support of why that particular answer was provided so that we could get buy-in and trust from our abstractors and drive efficiency from really 2 perspectives. One is not only getting the answer right, but taking the time away that they currently have to spend navigating into the EMR, finding where that relevant information is and getting all that upfront to them and in one place. The projected efficiency that we're seeing with this, this is ongoing, is around 24% less time spent abstracting. So schematically, what does this look like a patient is put into the queue to be abstracted before that patient is placed into the data entry UI, their information is first passed through our AI service. Now for each question within the registry, we first search for relevant documentation in answer to that question. That information is then passed into the generative model and interpreted. The first thing the model outputs is a recommended answer, and then it searches for justifications in support of that answer. The final output is what the recommended response is and then supporting evidence. Our model is fine-tuned to all of the data that we've collected over the last year within the different data entry applications that we have as well. So what does this actually look like live within the UI today. This is our ARMUS abstraction UI, the abstractor would select a particular question that they're looking to answer. In this case, we're looking at diabetes risk factors or diabetes control. Once they click on that particular question, our AI assistant over here updates. And as you can see, we have a recommended response and then the evidence in support of that response as well as a link out to the entire note itself. So they have all that information in 1 place to make an informed decision. In this case, the recommended answer was that the diabetes control was through oral medication and the support of that answer from the clinical text that was extracted was for them to continue the current treatment plan with metformin. Now as I mentioned, one of the things that we want to ensure that we're able to do is embed the abstractors expertise into this of the registry itself as well as the clinical components of it. So a little bit of a nuanced example here is with tobacco use. So in this particular case, abstractor again clicks on the particular field, our AI assistant updates. It recommends the answer that the patient is a former smoker, the reason that provides that response is that it's keyed in on a quick date, in this case, June 15, 2015. I promise there's no PHI here, by the way. However, according to the registry definition, tobacco use includes smokeless tobacco. In this case, this patient is using snuff. So while the AI assistant provided an incorrect recommendation, we surfaced the relevant information to the abstractor for them to be able to make a correct judgment call. For us, net-net, that's a win, because we've kept them out of having to navigate through the EMR in finding relevant information, and we've made the process for them simpler, consolidating their workflow. So I know we're up on time. I just want to say I'm incredibly excited for what we have coming down the pipe in Generative AI in 2024. I know there's a lot of hype, but I have to say that it's worth buying into especially if you're very deliberate and conscious with how you apply these tools. So thank you so much.

Great job. Thank you to all of our presenters. We really appreciate that opportunity. What we'd like to do now is a facilitated Q&A. And so we're going to open up the next 15 minutes or so, 15, 16 minutes until 7:30 and invite you to ask questions, and then I might just suggest who within our attendee group might be best equipped to provide thoughts and answers. And we've got roving mics. So let's get started. Dave, please.

So can you maybe talk a little bit about your decision to go with Health Catalyst like this is like in early 2020, you got COVID happening, massive decline in acute care volumes, inflation pressures like labor pressure, shortages. Like can you talk about that process? And then for 2023, it's been another difficult year. Where are you now in terms of like the margins at your hospital, how is your TEMS deal with Health Catalyst progressing? Are you achieving your milestones? And then like looking forward, like as analysts, we want to see that revenue growth ramp back up. My personal view is it was a tough year in 2023. Are things improving or not? Just any more color there would be very helpful.

Certainly. So I'll maybe stand up. So first part of the question, how did we select Health Catalyst to begin with? Was that sort of -- or...

Yes. Like why do you need the analytics if you're under such pressure with buying volumes, revenue pressure, inflation pressures. Why are you now spending money on technology platform and where are now in 2023?

Yes. Yes. So I think the reason that we're investing in this space is because in order for us to really truly understand the business better and understand some of the cost variables that we have to manage differently than we've ever had to manage. We need to have expertise to help us aggregate that information, but then also devise insights from it that we would maybe struggle to do on our own or at least at the same pace. I think what we've seen right now is historically, in healthcare, if you grew volume, typically revenue came with it. That's still true in some instances today, depending on the model. But cost management is becoming increasingly important with all the inflationary pressure that you mentioned, workforce, but also just low-value care and waste in the system. And without some of the insights that we've gleaned through this partnership, it's very difficult for us to manage that effectively. That's what I mentioned. That was really the goal at the outset back in 2019, was we have all these disparate data sources, but we're struggling ourselves to put them together in a way that helps to tell a story that be action-oriented. And as T.J. underlined, we want to drive data informed decision-making more proactively rather than react. And healthcare has done a good job at tracking information and looking backwards, but the real key is, I think, get that information and look forward. So that's a big part of the driver. Have we seen for Carle Health, and Dan mentioned some of the savings that TEMS will, we've certainly realized that. And the other thing for our organization over that 3-year period, we've added 5 hospitals in 2 service areas and been able to scale support for that integration that would not have been possible without this partnership.

Thanks, Matt. Jess?

Rarely hear that my voice is not up here saying enough. So I appreciate the microphone. I guess I just -- I kind of wanted to understand with the new platform, you guys are mentioning potentially Health Catalyst existing as a supplement or complement to other data management platforms. So just first off, I want to understand, is that new, what do you imagine the modernized kind of data architecture at a health system looking like? So right, everything from EHRs to other complements to HCAT to what exists on top of HCAT. And I'm just interested to know kind of cost of ownership prior platform versus new platform? How does that compare for your existing customers and for potentially new customers?

Yes, great questions. I'll share a few thoughts and then Dave, if you want to add to that. So it has always been the case, and Matt actually referenced this. And my goodness, the number of times we've seen clients that have invested in lots of different kind of competing data -- platforms data, warehouses data, whatever you want to call it, has been there for the 16 years of our existence. I think what's new and different. There are a couple of components, I think, are importantly new and different. One is that the technology capabilities coming out of Silicon Valley have accelerated. And so the benefits of being able to find an architecture that allows us to tap into that in a really meaningful way in the industry is more important. And if I contrast and Dave and T.J. mentioned this, our old DOS platform, one of the challenges we had 10 years ago, 15 years ago was, we had to build a lot of the componentry. It didn't exist. It wasn't coming out of Silicon Valley and tuned for healthcare use cases. Now a lot of that cross-industry capability has really caught up and there are amazing innovations that are coming out of Silicon Valley that those cross-industry technologies are now better, faster and cheaper than Health Catalyst having to build that componentry, those tools in a proprietary way. The next-gen data platform more fully takes advantage of those cross-industry capabilities. It does slim down that portion of the tech platform that we occupy, we're actually grateful for that. We have a necessity had to build some of that componentry, but it's better for someone else to build that at scale. And now we have vendors that are doing that. So that's better for the industry. It's better, faster and cheaper for the ecosystem for us to allow on the next-gen data platform tapping into that. I think the second thing I'll say and then I'll turn it to Dave is with the acceleration of innovation, including AI innovation and interesting companies, interesting capabilities coming out of cross-industry space, it becomes even more important to have a much more modular architecture. And that's where the next-gen data platform enables that for today but is also the right architecture for the future, just had several client discussions about the usefulness of that architecture that when the next OpenAI, the next interesting company that's bringing really cool technologies and capabilities comes to the fore. If you have this more flexible modular architecture, it's much easier to plug that in to the architecture that is part of the next-gen data platform, much harder to do that in the old systems and in the old approaches. And so I think increasingly, most health systems will want that more modular architecture because of the explosion in technology innovation, and it does lower the cost of the ecosystem cost. We are, just like Dave shared earlier in his presentation, able to scale down more because our cost of ownership is lower and still maintain meaningful margins. In fact, we expect margin expansion at the data platform layer, even as the price point comes down, we're able to serve more of the low end of the market. That's just better for the ecosystem. And we think this architecture not only takes advantage of that now is also tuned for the future. Dave, what would you add?

Dan, you covered it well. I think I love that we left the screenshot up because in environment where we did not have an enterprise-wide EDW DOS install a big bang solution, this would be very tough for us to do. But on the next generation platform, this can be a much more modular approach where we source the data that we need to do the AI pipelines to answer these questions in this just 1 solution, which is -- this is ARMUS, without needing the client to first fully commit and we love our clients to have to an enterprise-wide data solution. So we need that option, but we also need to have that full stack for them as well. So I think it's kind of like meeting where -- meeting them where they're at, where they can get started small with us and have a more modular relationship and then grow to something bigger as they realize we can provide much more of the stack. And yes, there are a large academic health sciences systems out there that they have multiple teams doing multiple different things. They have a research arm that's got its own data platform. They've got several affiliate hospitals that have built their own things. And it's really tough to just all in 1 dislodge all of those pieces. So it's better for us to think of that as a data ecosystem that we play in, then just there can only be 1 ring to rule them all and that's Health Catalyst. So it's kind of like a -- it's a sized approach. And the old generation platform didn't really give us the ability to offer that. We had kind of had to offer that just like all encompassing solution.

One other thing I would add, as I think back over the last 5 years and forward to the next 5 years and changes that I think, Jess, we will see, I shared that slide and Dave referenced it of the number of total clients quintupling with Health Catalyst and you saw a real distinction, there's 109 of those DOS subscription clients and then another 500 non-DOS. Today, there's a really big difference between the amount that a DOS client spends with us, it's multiple millions of dollars, right, on average, and the amount that a non-DOS client spends with us. It's a big step down, a hundred thousand -- a couple of hundred thousand. I think in the future, you're going to see the distance between those 2 filled, where there's going to be much more of a spectrum. To Dave's point, we're going to make it a lot easier to migrate from a couple hundred thousand to 500,000 to 800,000 to 1 million. And we can meet folks where they are, give them a much more natural path where in the past, it's been a huge kind of an order of magnitude change to go from 1 category to the other. It's going to be a lot easier and more fluid in the future. Let's go over here.

Daniel Grosslight of Citi. I want to go back to something that Matt mentioned, but maybe make it a little broader of a question. On building in shared incentives for clinical outcomes. I'm curious if you can give us some examples of what that might be and if that's kind of a newer contracting strategy that you're bringing to the market? And then secondly, as we think about product expansion within TEMS. On the ambulatory side, I noticed that denials management was one of the things you guys are doing, which brings into question, what more can you do within revenue cycle management, because that is obviously, one big area of focus for providers. So as we think through SKU expansion within TEMS, is RCM kind of top of mind there? Or what other products are you thinking of bringing to market next?

Yes. Great questions, Daniel. Thank you. I'll share a few thoughts and then maybe Dan LeSueur, if you want to be prepared with additional thoughts. So having shared incentives is something that we have experimented with our clients for over a decade. Our first TEMS relationship with Allina included a very meaningful shared incentive around outcomes improvements. We love that concept. We love the alignment that, that drives. There are some problems with those kinds of structures, especially around attribution, where part of what we all recognize is, gosh, everybody is contributing. There's a lot of factors that go into an outcome actually improving. And when you start attaching a lot of financial elements to it, then a lot of people start having significant opinions about that attribution and that can lead to some challenging partnership dynamics. And so we've ended up not doing a ton of that, except in one area in the emerging area of ambulatory operations that Daniel kind of walked through. We have found that, number one, most of our clients are struggling with their financial performance in the ambulatory operations. Number two, when we go in and start diagnosing with data, what the problems are, they're fairly equal parts revenue problems and cost problems. And so when we think about how we bring tech-enabled managed services to bear in a win-win-win, like Daniel described, actually having a bridge being a performance incentive is an important part because it TEMS in that SKU and ambulatory operations can't just be about cost savings. For it to be really positive and sustainable for us, we also need an opportunity to share in the success when revenue is enhanced, including through revenue cycle activities. And we do, Daniel, see some meaningful opportunities in the rev cycle space. We're also aware of how hard it is to be the best in the world at any number of things. So I think we're trying to have humility and recognizing maybe there's a few aspects of revenue cycle management, like mid rev cycle charge master management. We do believe we're the best in the world of that with our Vita Water solution. And we do believe and we have ambition in the ambulatory operations space to help with specific use cases around denials management because that matters a lot in ambulatory operations. So we are going to focus there and you may see us do some things both from a technology development perspective and a services perspective, but I think it will be more narrow. Daniel, what would you add?

Yes, just to speak to the ambulatory example that Dan was talking about, it's much more than just driving efficiencies or cost savings, right? It's also enhancing revenue opportunities. And so the metric that we zeroed in on within that -- those engagements is net operating income. So whatever we can do to improve net operating income for this book of business in the ambulatory space for our customer, there's a share of that improvement that we claim. The vast majority of that improvement goes to the client, but there is a portion, a sliver of that improvement that comes to us as we improve that. And this keeps us focused on the whole picture, not just being -- not just reducing head count and squeezing blood out of the turnip, so to speak, but really growing the business, extending their or improving patient access, improving their performance in contracts, et cetera.

Thanks, Daniel. Maybe time for one more question. David, let's go to you.

I think one of the most difficult things for people to conceptualize in the TEMS businesses starting at 0% gross margin and then ending up at 25%. So that's a big gap. Maybe you could just talk a little bit -- what are some of the milestones that you need to achieve along that path to get to 25% while still delivering all the cost dynamics and even some of the revenue dynamics that were just discussed to kind of hit that objective?

Yes, it's a great question, David. And it isn't easy. And I think that is actually one of the reasons why we're excited about it. It's hard to do. It takes both technology components that automate steps in the process that were manual. We're constantly looking for new technology-related elements that we can bring to bear. And Dan just walked -- Dr. Dan just walked us through an example of that for charter extraction that isn't embedded in our forecast, by the way, of getting from 0% to 25% that additional 24% efficiency gain, first of all, is early. So don't build it into your forecast, please, yet. But that would be above and beyond what we've already demonstrated as our capacity and capability. And that 0% to 25% gross margin is data informed in terms of what we already achieved with the technologies that we've already purchased or developed, like ARMUS, like ERS, like the data platform, but we're constantly looking for more. And it is really important that TEMS starts with the T right, that the technology enablement is a really, really big deal. The last thing I'll share and then, Daniel, if you'd like to add anything or Dr. Dan, if you like to add anything, is that as we move forward and deepen the relationship like what Matt Kolb described, when we enter into a tech-enablement services relationship, in each case, in the vast majority of cases, we're also expanding the tech subscription. So we're broadening tech footprint at those clients, that has a very different starting gross margin, right, more in the 70-plus percent range. And so that contributes to the overall health of that client relationship while we're ramping up the tech-enablement services components. Daniel, anything you'd add?

I was just going to mention that the slide says this, but I didn't call it out earlier, that as we -- there's not a lot of additional OpEx that goes with expanding into these TEMS relationships. And so when we solve the equation for gross margin, almost all of that drops to EBITDA. And so we're able to rinse and repeat the plays that we've run with the early clients in each subsequent client. And we're already starting to see a lot of that leverage in chart abstraction as well as in data analytics, and we're building that library now in the ambulatory space and some other emerging areas.

I'd just add that one of the things that TEMS kind of uniquely positions us to be able to do from an innovation perspective is that we have an internal audience that we're building toward we have abstractors that are our own team members. We have analysts that we're bringing on from Carle who we can build these services and tools against and iterate rapidly. So we have a friendly audience that's going to provide us detailed feedback, and we know that we can build the best solution to our team members, and as we refine and harden around those solutions, we can open them up to the broader community of Health Catalyst customers. So I would say that there's this interplay between our ability to innovate more rapidly as a result of TEMS and it's kind of an interesting place to be.

It's a great point and maybe a last comment. When we are managing the team, we found this across every TEMS engagement, every TEMS relationship. When the team is with Health Catalyst, our ability to accelerate adoption of new technologies and accelerate the adoption of process improvements. One of the most important slides that you saw tonight was Allie Corona's slide with all those boxes, at 1 level, it looks like, oh, you only get improved by 5 minutes. But remember what she said, that's -- multiply that by hundreds of thousands of cases, right? 5 minutes per case matters a lot. And Allie is one of our best leaders at Health Catalyst. And to Dr. Dan's point, she manages that organization in such a way that everyone wakes up every morning focused on how do I find 30 seconds of improvement. And that can include because I love technology, and I'm going to adopt technology that can include because I'm going to share learnings with my colleagues, I'm going to figure out process improvements. And it isn't any 1 thing. It's hard to do. But that's also why we believe we can be the best in the world at it. We're committed to it. We're maniacally focused on it. We're good at it, and we're just getting started. So maybe with that, we'll close the formal Q&A. But a number of us can stick around for a little while. We really appreciate your time. It's been great to be together. Thank you very much.

And just -- sorry, just a quick reminder for -- as Dan mentioned, as many of us as possible, we'll stay here. And then for anyone else who would like to go downstairs to the first floor, our analytics showcase and AI showcase are going on the first floor of this hotel. So if you want to find your way down there. As I mentioned, good opportunity to interact with clients and see some of the success stories from our clients' mouths. So thank you again. Appreciate it.